Archive · 22 Aug 2026
AI, Human Agency, and Who Controls Algorithmic Conditioning
People do not create their ideas completely by themselves.
Families, schools, culture, media, and communities influence how people think.
The article says AI is entering this already-shaped world.
AI could help more people learn, ask questions, and think clearly.
This could help people who have had less access to good educational resources.
But AI systems are also built by companies and institutions with their own goals.
Personalized recommendations may show people more of what they already like and fewer new or challenging ideas.
Over time, this can influence what people pay attention to and believe.
The author says society needs to decide who controls these systems and make sure they serve many people rather than only powerful organizations.
Human agency is real but shaped by language, culture, institutions, media, and social conditions.
The article argues AI can act as a reasoning companion that expands access to learning and intellectual participation.
Unlike older information technologies, AI can help people interpret, compare, summarize, and reason with information.
Personalized algorithms may reinforce existing preferences, narrow exposure to disagreement, and shape attention without obvious coercion.
The author calls for transparency, diverse datasets, language inclusion, and public accountability over AI systems.
- Who
- The article’s author, Sreejith Sreedharan, discusses individuals, AI developers, companies, governments, and the public.
- What
- An opinion analysis of how AI may expand reasoning and learning while also conditioning thought through concentrated, personalized algorithms.
- When
- In the current period, as AI systems increasingly shape public life.
- Where
- Across public life and global knowledge systems, including examples involving learners in Odisha and Oaxaca.
- Why
- To argue that AI governance, transparency, inclusion, and accountability are necessary because algorithms influence what people see, repeat, and accept.
AI as Cognitive Levelling
AI as Concentrated Influence
Effect on human capability
AI as Cognitive Levelling
AI could raise the cognitive floor by helping more people reason, learn, write, and participate in intellectual life.
AI as Concentrated Influence
AI may strengthen existing inequalities if its systems are designed mainly to serve dominant institutions, profitable markets, or a small group of powerful firms.
Personalization
AI as Cognitive Levelling
Adaptive systems can provide convenient, individualized support and make intellectual assistance more widely available.
AI as Concentrated Influence
Personalization can repeatedly return familiar preferences, narrow exposure to different ideas, and shape attention without visible coercion.
Governance
AI as Cognitive Levelling
With inclusive design and public accountability, AI could support plural thought and broader human development.
AI as Concentrated Influence
Without strong oversight, corporate or government control could consolidate intellectual power and encode narrow cultural and political priorities at scale.
- Central argument
- The key issue is not whether human agency was ever completely independent, but who governs the algorithmic conditioning that now shapes thought.
- Potential benefit
- AI can explain, translate, compare, summarize, test assumptions, and support learning.
- Access
- The article says AI could lower barriers created by geography, class, language, and institutional access.
- Main concern
- Frontier AI is concentrated among a relatively small number of firms and research ecosystems.
- Algorithmic effect
- Personalization can reinforce existing preferences and reduce exposure to disagreement and surprise.
- Cultural risks
- English-language dominance, Western institutional habits, and commercial incentives may influence AI responses and recommendations.
- Proposed safeguards
- The author calls for transparency, language inclusion, dataset diversity, accountability, and greater public oversight.